[flaky-test] fix duplication check after introducing skips (#78958)
My last change #78292 caused a regression in print test stats, which was entirely my fault. e.g., https://github.com/pytorch/pytorch/runs/6750407899
The eventual goal is to retire the flaky test calculations here, but for now let's keep reporting them before switching over our infra to the newer flaky test aggregation
Pull Request resolved: https://github.com/pytorch/pytorch/pull/78958
Approved by: https://github.com/suo
